The efficacy of treatment for non-small cell lung cancer (NSCLC) has improved only marginally over the past few decades. Chemoradiotherapy continues to have a narrow therapeutic index, and response rates are unpredictable, even after controlling for the known predictors of stage, histology, grade, and performance status. Drug resistance presents a major obstacle in NSCLC treatment. Platinum agents, specifically cisplatin and carboplatin, are the cornerstones of NSCLC therapy. Radiation induced toxicity is similarly unpredictable and dose limiting. Data from in vitro studies show that radioresistance is associated with poor prognosis and that, compared to radioresistant lines, radiosensitive cell lines exhibit greater initial damage and repair this damage more slowly. Cellular events that are modulated by sets of interacting genes may partly explain such unpredictable outcomes. We hypothesized that interindividual genetic differences impact the efficacy and toxicity of platinum-based therapy and that this genetic information can be used to predict individual drug response. In this proposal, we will build upon an existing case ascertainment infrastructure to achieve the following specific aims: 1. We will construct a well-characterized cohort of 1400 advanced NSCLC (inoperable stages III and IV) patients treated with platinum drugs combined with other agents + radiation. This sample will consist of 800 patients prospectively enrolled and 600 patients retrospectively identified through participation in our ongoing studies. By personal interview and medical chart review, we have obtained or will obtain detailed baseline information on smoking history, dietary intake, family cancer history, previous medical history, weight loss, performance status, tumor characteristics, treatment toxicity and efficacy, and complete patient follow up data. 2. To evaluate the intrinsic repair capacity of the prospective cases, we will use two functional lymphocyte-based comet assays, the fir it of which assesses basic excision repair (BER) and double-strand break repair (DSB) of radiation-induced damage, and the second of winch assesses mainly nucleotide excision repair (NER) of platinum-induced damage. Through these assays, we will test the hypothesis that the host's intrinsic repair capacity influences drug sensitivity. Specifically, we hypothesize that reduced NER capacity is associated with improved response to platinum therapy (due to decreased ability to repair platinum adducts). Further, we hypothesize that reduced BER and/or DSB repair capacity, because it translates to a decreased ability to repair damage to rapidly dividing cells, is associated with higher risk of acute and late-term radiation toxicity;however, reduced BER and/or DSB repair capacity is conversely associated with improved survival. 3. In parallel with the phenotypic assays, we will assume a pathway-based approach in all 1400 patients to evaluate frequencies of polymorphisms in genes involved in DNA repair pathways, specifically NER, BER, and DSB. We will also evaluate frequencies of polymorphisms in pathways relevant for the activity or disposition of platinum analogs and taxols, including drug metabolism, multidrug resistance, cellular uptake, and select genes in cell cycle control and apoptotic pathways. We will test the hypothesis that therapy response, resistance, and toxicity are modulated by variants in multiple relevant genes. We will also test the hypothesis that there are correlations between the repair genotypes and the phenotype data from Aim 2. 4. We will integrate clinical and epidemiologic data - such as smoking at diagnosis and treatment initiation, demographic factors, poor dietary intake of select micronutrients, weight loss, performance status, and tumor characteristics - with the genetic data from the studies described above. We will develop a quantitative multivariate risk assessment model that will incorporate and prioritize these relevant epidemiologic and molecular predictors of outcome. The unifying theme of this study is that intrinsic differences in the susceptibility of cells to the genotoxic effects of radiotherapy and chemotherapy modulate toxicity and outcome. This multivariate prognostic study could have long-term prognostic potential by allowing greater individualization of therapy, based on the genetic makeup of the patient, thereby both increasing efficacy and reducing morbidity.